I have become comfortable with the JMP (from SAS) tools for using design of experiments (DOE). I share here some compelling statements from the JMP website:
ttps://www.jmp.com/en_ph/statistics-knowledge-portal/what-is-design-of-experiments.html
“What is design of experiments?
“Design of experiments (DOE) is a systematic, efficient method that enables scientists and engineers to study the relationship between multiple input variables (aka factors) and key output variables (aka responses). It is a structured approach for collecting data and making discoveries.
“When to use DOE?
- To determine whether a factor, or a collection of factors, has an effect on the response.
- To determine whether factors interact in their effect on the response.
- To model the behavior of the response as a function of the factors.
- To Optimize the response.
“Ronald Fisher first introduced four enduring principles of DOE in 1926: the factorial principle, randomisation, replication and blocking. Generating and analysing these designs relied primarily on hand calculation in the past; until recently practitioners started using computer-generated designs for a more effective and efficient DOE.”
The quotation from the stated website (from JMP) gives the basis of the arguments that I am asserting about why DOE is such a valuable tool for insect rearing personnel. First, the idea of studying the relationships between multiple factors should appeal to every rearing specialist. When we put our insects into a rearing system and provide them with artificial diets, we should wonder how the diet components affect the insect and how they affect one-another. For example, how does the pH of the diet relate to solubility of a protein supplement such as casein? Furthermore, how does the pH affect solubility of salt mixtures such as Wesson salts? A little inquiry about casein (which should more properly be called “caseins” because there are several milk proteins that are indicated by the term “casein”) shows that this protein is very hard to dissolve at near neutral pH levels. Therefore, unless we make special efforts to dissolve casein (such as pre-dissolution in basic solutions), this important protein is likely to form pockets within the diet where it is non-homogeneously distributed. The casein-rich vs. casein-poor pockets contribute to the problem of non-uniformity, which leads to serious differences in growth and development rates of larvae that are otherwise similar in age and genetic composition. The same type of non-homogeneity applies to dissolution and distribution of salt mixtures. I have shown (Cohen 2015) that Wesson salts are virtually non-soluble unless they are dissolved in a low pH solution. That helps explain some of the positive effects of using diets that are on the acidic side, rather than neutral. Further analysis helps us understand that the texture of the diet, in terms of firmness or gel strength, is enhanced by the presence of salt mixtures where Wesson and Beck salts have an enhancing effect on the gel strength of agar, carrageenan, or other hydrocolloid gels (Cohen unpublished data).
In Figure 1, I have tried to suggest some of the complicated interactions that are present in an insect diet. The lesson here is that using DOE-based investigations, we can begin to dissect the components of our rearing systems to help us understand the ways that rearing factors affect the nature and success of our rearing systems. The whole thrust of this approach is to turn the “black boxes” that describe our rearing systems into more transparent and logical processes.
More about this in future postings!